This is the program for the 2010 Joint Statistical Meetings in Vancouver, British Columbia.

Abstract Details

Activity Number: 406
Type: Contributed
Date/Time: Tuesday, August 3, 2010 : 2:00 PM to 3:50 PM
Sponsor: Biometrics Section
Abstract - #307407
Title: The Role of Proxy Genes in Predictive Models: An Application to Early Detection of Prostate Cancer
Author(s): Jay Magidson and William Oh and Robert Ross and Philip Kantoff and Karl Wassmann*+
Companies: Statistical Innovations Inc. and Mount Sinai School of Medicine and Infinity Pharmaceuticals and Dana-Farber Cancer Institute and Source MDx
Address: 199 Wells Ave. , Newton Center, MA, 02459,
Keywords: gene expression ; correlated component regression ; high dimensional data ; variable selection ; logistic regression ; proxy gene

The most important predictors often do not predict the outcome variable directly, but enhance the effects of other predictors in the model. The most important predictor in a 9-gene model for early detection of prostate cancer is SP1, a proxy gene, whose mean expression does not differ between cancer and normal subjects. However, SP1 is predictive of the pre-cancer expression of genes involved in its cellular process interactions, including some whose expression changed with cancer onset (prime genes). By subtracting out the predicted pre-cancer expression from the prime genes, the prime gene effects transform into stronger effects associated with post-pre change in prime gene expression. We present new models and methods that identify proxy and associated prime genes for inclusion in the model, from a potentially large pool of gene candidates. These methods generalize to survival models.

The address information is for the authors that have a + after their name.
Authors who are presenting talks have a * after their name.

Back to the full JSM 2010 program

2010 JSM Online Program Home

For information, contact or phone (888) 231-3473.

If you have questions about the Continuing Education program, please contact the Education Department.